Beyond Peer Influence and Selection: A Gendered Pathways Analysis of Social Environmental Predictors of Change in Child and Peer Delinquency

Abstract

The purpose of this study was to determine whether maternal closeness and participation in unstructured routine activities differentially predicted change in child and peer delinquency for female and male youth above and beyond the effects of peer influence and selection. Participants were 3,370 (1,759 boys, 1,611 girls) members of the Fragile Families and Child Welfare Study. When regression analyses were performed on boys and girls separately, unstructured routine activities effectively predicted a rise in child and peer delinquency in boys and maternal closeness successfully predicted a drop in child and peer delinquency in girls, findings consistent with gendered pathways theory.

Keywords

peer influence peer selection closeness to mother unstructured routine activities

Criminologists have traditionally advanced two principal interpretations of the well-known relationship between child and peer delinquency. They are commonly referred to as the peer influence and peer selection effects. The peer influence effect assumes that the individual learns the attitudes, behaviors, and techniques necessary for crime through his or her associations with those already involved in crime. This position is adopted by theorists affiliated with social learning school of criminology (Akers, 1998; Warr, 2002). The peer selection effect is said to occur when someone already involved in crime begins associating with a delinquent peer group through shared interests, values, and beliefs. This homophily or “birds of a feather” perspective on the child–peer delinquency relationship is adopted by proponents of social and self-control theories of crime (Gottfredson & Hirschi, 1990; Kiesner, Poulin, & Nicotra, 2003). The peer influence and peer selection effects have both received strong empirical support, and it is generally agreed that each plays a role in explaining the well-documented connection between child or participant delinquency and peer delinquency (Monahan, Steinberg, & Cauffman, 2009; Sedding, 2014; Svensson, Burk, Stattin, & Kerr, 2012). The unique contribution this study makes to the research literature on delinquency and gender is testing the possibility that two variables differentially explain future peer and child delinquency in males and females above and beyond the effects of peer influence and selection. Two social variables with potentially divergent effects on girls and boys, that is, maternal closeness and unsupervised routine activities, served as predictor variables in this study, and a third variable, parental knowledge, served as a control variable, given its potential impact on the predictor variables.

Maternal Closeness

Two putative predictors were selected for analysis in this study based on prior research showing that each does a reasonably good job of predicting child delinquency, peer delinquency, or both. Emotional attachment or closeness to a parent is one such predictor. Using longitudinal data from the Gang Resistance Education and Training (GREAT) study, Higgins, Jennings, and Mahoney (2010) uncovered a strong negative correlation between estimates of maternal and paternal bonding, on one hand, and criminal offending, on the other hand. Four years later, Schroeder, Higgins, and Mowen (2014) discovered that although maternal attachment in adolescence correlated negatively with self-reported offending, the effect was no longer significant once these individuals had entered early adulthood. More recently, Alvarez-Rivera (2016) determined that maternal attachment correlated with self-reported criminality in a group of White and Hispanic college students, independent of the effects of low self-control. In a study that examined maternal and paternal closeness and control, Vieno, Nation, Pastore, and Santinello (2009) observed a significant effect for maternal closeness and control that varied as a function of gender: Specifically, maternal control predicted delinquent behavior in boys but not girls, and maternal closeness predicted parental knowledge in girls but not boys. Although maternal closeness predicted delinquency in both boys and girls, the effect was more than twice as strong in girls.

Unstructured Routine Activities

Unlike maternal closeness, which is classified as a protective factor because it decreases a child’s chances of future delinquent involvement, unstructured routine activities are classified as risk factors, in that, they increase a child’s odds of future delinquent involvement (Osgood, Wilson, O’Malley, Bachman, & Johnston, 1996). Analyzing data on Chicago youth gangs gathered between 1959 and 1962, Hughes and Short (2014) observed a link between immersion in unstructured routine activities and involvement in delinquent behavior in many of these gang members. Based on the results of this study, Hughes and Short concluded that gangs contribute to delinquency, in part, by providing youth with increased opportunities for unsupervised socializing with peers. Riding around in cars, hanging out at malls, and congregating on street corners are all ways juveniles increase their opportunities for engaging in negative behavior. There is every indication that participation in unsupervised activities with peers predicts child delinquency after accounting for peer deviance (McGloin & Shermer, 2009) and that the effect travels along multiple paths, some of which do not include peer delinquency (Hoeben & Weerman, 2016). Studies by Novak and Crawford (2010) and Augustyn and McGloin (2013) demonstrate that participation in unstructured and unsupervised activities with peers is more criminogenic in boys than in girls.

Parental Knowledge

Research suggests that perceived parental knowledge may have a stronger impact on future delinquent behavior than parental supervision or surveillance (Kerr & Stattin, 2000), and that it may achieve its counter-delinquency effect by reducing a child’s involvement in unsupervised routine activities (Walters, 2018b) or improving a child’s feelings toward his or her parents (Walters, 2018a). A third possibility surfaced in a study by Cutrin, Gómez-Fraguela, and Sobral (2017), whereby parental support and closeness affected delinquency indirectly by increasing parental knowledge. In a study surveying the relationship between perceived parental knowledge and peer delinquency in Chinese immigrant families, Wang, Kim, Anderson, Chen, and Yan (2012) discovered that greater acculturation discrepancies in parent–child dyads were associated with decreased parental knowledge and increased contact with deviant peers. Previously, Reitz, Prinzie, Dekovíc, and Buist (2007) had discerned that decreased parental knowledge and a poorer quality parent–child relationship coalesced, to where peer delinquency and involvement in unstructured routine activities with peers created a rise in youth involvement in various forms of delinquent behavior. Given the fact that perceived parental knowledge may interfere with maternal support and promote unstructured routine activities, it was treated as a control variable in the current investigation.

Gendered Pathways Theory

The conceptual framework for the current study was provided by gendered pathways theory. According to gendered pathways theory, the routes males and females take to crime are similar in some ways and different in others. Although founded on qualitative and descriptive research (Belknap, 2007; Byrne & Trew, 2008; Simpson, Yahner, & Dugan, 2008), gendered pathways theory has become the subject of an ever-increasing number of quantitative studies (Kruttschnitt, 2013). Using a growth-based person-centered analytic approach, Brennan, Breitenbach, Dieterich, Salisbury, and Van Voorhis (2012) identified multiple pathways of criminal involvement in a sample of female offenders. Some of these pathways mimicked trajectories observed in male offenders (Moffitt, 1993), whereas others replicated patterns first observed in qualitative studies on female offenders (Daly, 1992). A measure such as the Level of Service Inventory–Revised (LSI-R: Andrews & Bonta, 1995), which is an effective risk assessment technique for use with male offenders, has been found to be of limited value in working with female offenders, except in those cases where the female offender displayed a male-like pattern of offending (Reisig, Holtfreter, & Morash, 2006). By contrast, variables such as low self-control and risky lifestyles appear to be equally applicable to male and female violent victimization (Turanovic, Reisig, & Pratt, 2015).

Quantitative research indicates that early physical and sexual maltreatment, family relationship issues, substance misuse, and mental health problems play a more salient role in female offending than in male offending (Belknap & Holsinger, 2006; Chesney-Lind & Palko, 2004). Several of these factors were recently incorporated into a single gendered pathway in which early sexual abuse led to internalizing symptoms, which then led to drug use, and which culminated in criminal offending (Broidy, Payne, & Piquero, 2018). A principal limitation of quantitative research on gendered pathways theory, however, is that much of it is based on the retrospective accounts of incarcerated female adolescent (Jones, Brown, Wanamaker, & Greiner, 2014) and adult (Salisbury & Van Voorhis, 2009) offenders; and, when real-time analyses of nonincarcerated youth are undertaken, the samples have often been restricted to female respondents (Bloom, Owen, Rosenbaum, & Deschenes, 2003). Of the handful of studies that have prospectively examined gendered effects in nonincarcerated male and female youth, the most consistent findings are for family and peer factors. Using data from the National Longitudinal Survey of Youth–Child, for instance, Walters (2013) determined that parental involvement linked early delinquency to later criminality in girls but not boys, whereas perceived peer pressure mediated the relationship between early externalizing behavior and subsequent delinquency in boys but not girls (Walters, 2014).

The Current Investigation

This study sought to contribute to the gendered pathways literature by testing whether maternal closeness and unstructured routine activities predicted future delinquency above and beyond the peer influence and selection effects in youth making the transition from childhood to adolescence. Based on prior research showing that sex is capable of moderating the effect of these social environmental factors on delinquency, it was argued that although the peer influence and selection effects would be observed in both male and female youth, maternal closeness would do a significantly better job of predicting child and peer delinquency in girls than in boys, whereas involvement in unstructured and unsupervised routine activities would do a significantly better job of predicting child and peer delinquency in boys than in girls. A quantitative, rather than qualitative, model was proposed such that boys, like girls, could potentially be protected by maternal closeness, and girls, like boys, could potentially be placed at risk by unstructured routine activities; it is just that the former effect will be significantly stronger in girls and the latter effect significantly stronger in boys.

Two hypotheses were tested in this study. The first hypothesis held the following:

Hypothesis 1: Peer influence and peer selection effects would be significant in boys as well as girls.

The second hypothesis predicted the following:

Hypothesis 2: Despite significant peer influence and selection effects, maternal closeness would exert a significantly stronger inhibitory effect on child and peer delinquency in girls than in boys, and unstructured routine activities would have a significantly stronger facilitative effect on child and peer delinquency in boys than in girls after controlling for the effects of parental knowledge.

Given the fact that the quantitative superiority of maternal closeness in girls and unstructured routine activities in boys was predicted a priori, one-tailed z tests were used to evaluate the significance of any differences found between boys and girls in this regard.

Method

Participants

Participants for this study were members of the longitudinal Fragile Families and Child Wellbeing Study (FFCWS: Reichman, Teitler, Garfinkel, & McLanahan, 2001). The FFCWS is a nationally representative sample of 4,898 children born between 1998 and 2000 in U.S. cities with populations of 200,000 or more. Using disproportionate stratified random sampling, Reichman et al. (2001) oversampled nonmarital births for the purpose of creating a large pool of “at-risk” families. The child’s mother and, where possible, father were interviewed/surveyed prior to the child’s birth and along with the child and the child’s teachers, reinterviewed/resurveyed in five follow-up evaluations occurring when the child was 1 year of age (Wave 2), 3 years of age (Wave 3), 5 years of age (Wave 4), 9 years of age (Wave 5), and 15 years of age (Wave 6).

Children with complete data on at least half of the 10 variables used in this study (including sex) were retained as study participants. The present sample accordingly consisted of 1,759 boys and 1,611 girls, all of whom were 9 years of age at Wave 5 and 15 years of age at Wave 6. Comparing the 3,370 children who participated in the current study with the 1,528 children who were left out because of excessive missing data, produced six variables on which at least 10 nonparticipants had complete data (sex, race, household income, peer delinquency-9, child delinquency-15, and peer delinquency-15). A series of t tests revealed that nonparticipants included significantly more White children from homes with higher household incomes (p < .05).

The racial/ethnic breakdown of the total sample (N = 3,370) was 17.3% White, 50.0% African American, 24.9% Hispanic, 5.3% multiracial, and 2.5% Other. Nearly 90% of the children in the FFCWS were living with their biological mothers at the time of the Wave 6 interview and only about a quarter (28.6%) of the biological mothers were still married to or romantically involved with the child’s father at the time of the interview. This was one reason why closeness to the biological mother was included as a predictor in this study rather than closeness to both parents. The median annual household income for the families of FFCWS children was US$45,000 at Wave 6.

Measures

Demographic and control measures

Three control variables were included in the current investigation: race (White = 1, non-White = 2), annual household income (in thousands of dollars), and perceived parental knowledge. Perceived parental knowledge was evaluated with a single item: “how often does your principal guardian (usually the biological mother) know which friends you hang out with.” The child rated this item on a 4-point scale (0 = never, 1 = sometimes, 2 = often, 3 = always). All control variables were assessed during Wave 5 (age 9 years) interview.

Child delinquency

Participant or child delinquency was assessed when participants were 9 and 15 years of age. At age 9, participants were asked whether they had engaged in any of the following 13 antisocial acts (damaged property, stole something, took money from home, cheated on a school test, fist fight, trespassed, ran away from home, skipped school, suspended from school, graffiti, set fire to a building or car, slipped into a movie or sporting event without paying, and threw rocks or bottles at cars or people). Youth received 1 point for each antisocial act engaged in over the past year to create a scale that ranged from 0 to 13. At age 15 participants were asked whether they had participated in any of the following 12 acts (graffiti, property damage, shoplifting, fighting, hurting someone bad enough they needed medical treatment, auto theft, stealing something worth less than US$50, stealing something worth more than US$50, burglary, robbery, drug sales, and participation in a group fight) in the last year. Each offense was rated on a 4-point scale (1 = never, 2 = 1 or 2 times, 3 = 3 or 4 times, 4 = 5 or more times) to yield a scale that could range from 12 to 48.

Peer delinquency

Peer delinquency was also measured at ages 9 and 15. At age 9, peer delinquency was assessed by the primary caregiver (usually the biological mother), and at age 15, it was assessed by the child. During Wave 5 of the FFCWS, when the children were 9 years of age, mothers were asked to evaluate the following statement: “the child hangs around with others who get in trouble,” on a 3-point scale (1 = not true, 2 = somewhat or sometimes true, 3 = very true or often true). Peer delinquency at age 15 years was rated by the child using a series of five items (“friends deliberately damaged property that did not belong to them,” “friends stole something worth less than US$50,” “friends stole something worth more than US$50,” “friends used or threatened to use a weapon to get something,” “friends sold marijuana or other drugs”), each of which was rated for the past year on a 3-point frequency scale (1 = never, 2 = sometimes, 3 = often). This produced a score that could range from 5 to 15, and which achieved adequate internal consistency in the current sample of participants (α = .76).

Predictors

Two predictors were included in this study. These variables served as potential facilitators or inhibitors of change in child and peer delinquency. Each was assessed with a single item administered at Wave 5 of the FFCWS when the child was 9 years of age. Closeness to the (biological) mother was assessed with a single item that asked the child to rate “how close do you feel to your mom” on a 4-point scale (1 = not very close, 2 = fairly close, 3 = quite close, 4 = extremely close). Degree of involvement in unstructured routine activities with friends was also assessed with a single item, in which the child was instructed to estimate how much time he or she spent hanging out with friends on a typical weekday (0 = none, 1 = half an hour or less, 2 = more than half an hour but less than an hour, 3 = 1-2 hr, 4 = more than 2 hr).

Research Design and Statistical Analysis

The research design for this study consisted of two models: a peer influence model (peer delinquency at age 9 → child delinquency at age 15) and a peer selection model (child delinquency at age 9 → peer delinquency at age 15). To achieve maximum confidence in the temporal direction of the variables, given that participants were not randomly assigned to conditions, a prior measure of each predicted variable (i.e., child delinquency at age 9 in the model predicting child delinquency at age 15 and peer delinquency at age 9 in the model predicting peer delinquency at age 15) was included in the analyses (Cole & Maxwell, 2003). In this way, the predictor variables were actually measuring a change in child or peer delinquency from age 9 to age 15 rather than a static estimate of child or peer delinquency at age 15. Two multiple regression equations, one of which predicted child delinquency at age 15 and the other of which predicted peer delinquency at age 15, were computed separately for male and female participants using Mplus 8.1 (Muthén & Muthén, 1997-2017). A maximum likelihood with robust standard errors (MLR) estimator was employed in all analyses.

Missing Data

Complete data were available for 57.9% of the male sample (1,018 out of 1,759), with another 26.8% of males missing data on one variable and 15.4% missing data on two to five variables. Similarly, complete data were available for 60.4% of the female participants (973 out of 1,611), with another 27.4% of females missing data on one variable and 12.3% missing data on two to five variables. Five variables had more than 10% missing data: female household income (29.5%), male household income (28.1%), male race (15.0%), female race (11.2%), and female peer delinquency at age 15 (10.6%). When children with complete data were compared to children with missing data, there were two out of 18 significant group differences, both in the male sample: Male children with missing data lived in homes with lower household incomes and were rated by their mothers as higher in Wave 5 peer delinquency than male children with complete data (p < .05).

Missing data were handled in this study with full information maximum likelihood (FIML). Rather than imputing missing values or removing cases with missing data, FIML calculates model parameters and standard errors based on all available data, and then applies the estimated model parameters and standard errors to the entire sample of participants, FIML generates estimates that are more accurate and significantly less biased than those produced by more traditional missing data procedures such as simple imputation, listwise deletion, and pairwise deletion (Allison, 2012; Peyre, Leplége, & Coste, 2011). Research indicates that FIML is robust to violations of its basic assumptions, and that its results are trustworthy under all but the most extreme circumstances (Collins, Schafer, & Karn, 2001).

Results

Preliminary Analyses

Tables 1 and 2 list the means, standard deviations, ranges, and zero-order correlations for the nine variables included in this study, for the male and female samples, respectively. Nearly two thirds of the zero-order intercorrelations in both tables achieved significance using a Bonferroni-corrected alpha level. There was no evidence of multicollinearity between predictors in either the male or female regression analyses (tolerance = .933-.989, variance inflation factor [VIF] = 1.011-1.072). Because the outcome measures used in this study were moderately to highly nonnormal (skew = 2.69-4.25, kurtosis = 8.39-25.16), an MLR estimator was used instead of the standard ML estimator.

Table 1.

Descriptive Statistics and Correlations for the Nine Predictor, Outcome, and Control Variables: Male Participants.

Variable	n	M	SD	Range	2	3	4	5	6	7	8	9
1. Race	1,495	1.83	—	1-2	−.32^†	.01	−.04	−.06	.05	.09^†	.11^†	.11^†
2. Household income	1,265	48.52	56.36	0-900		.04	.07	−.02	−.10^†	−.08	−.14^†	−.13^†
3. Parental knowledge	1,739	2.13	1.06	0-3			.08^†	.10^†	.01	−.03	−.07	−.02
4. Close to mother	1,701	3.61	0.75	1-4				.04	−.03	−.03	−.10^†	−.02
5. Routine activities	1,725	2.14	1.48	0-4					.05	.07	.09^†	.08^†
6. Peer delinquency-9	1,728	1.15	0.39	1-3						.09^†	.20^†	.13^†
7. Peer delinquency-15	1,572	5.74	1.50	5-15							.13^†	.53^†
8. Child delinquency-9	1,742	1.47	1.82	0-13								.23^†
9. Child delinquency-15	1,583	13.35	2.59	12-36

Note. Variable = study variables; n = number of participants with nonmissing data; range = range of scores in the current sample; race = 1 (White) or 2 (non-White); household income = household income in thousands of dollars when child was 9 years of age; parental knowledge = child’s evaluation of parental knowledge of peer associations at age 9; close to mother = child’s rated closeness to mother at age 9; routine activities = unsupervised routine activities with peers assessed at age 9; peer delinquency-9 = peer delinquency at age 9 as rated by mother; peer delinquency-15 = peer delinquency at age 15 as rated by the child; child delinquency-9 = self-report of child’s own delinquency at age 9; child delinquency-15 = self-report of child’s own delinquency at age 15.

†

p < .0014 (Bonferroni-corrected α level; .05 / 36 correlations).

Table 2.

Descriptive Statistics and Correlations for the Nine Predictor, Outcome, and Control Variables: Female Participants.

Variable	n	M	SD	Range	2	3	4	5	6	7	8	9
1. Race	1,430	1.83	—	1-2	−.36^†	.00	−.04	−.04	.08	.06	.05	.09^†
2. Household income	1,135	45.61	48.40	0-600		.00	.06	.03	−.06	−.08	−.10^†	−.14^†
3. Parental knowledge	1,593	2.25	1.00	0-3			.12^†	.10^†	.00	.04	−.07	.02
4. Close to mother	1,567	3.66	0.72	1-4				.03	−.08	−.07	−.08	−.10^†
5. Routine activities	1,594	1.99	1.51	0-4					.03	.04	.07	.03
6. Peer delinquency-9	1,573	1.09	0.30	1-3						.11^†	.18^†	.12^†
7. Peer delinquency-15	1,479	5.58	1.31	5-15							.19^†	.46^†
8. Child delinquency-9	1,602	0.81	1.30	0-10								.21^†
9. Child delinquency-15	1,488	12.80	1.84	12-32

Note. Variable = study variables; n = number of participants with non-missing data; range = range of scores in the current sample; race = 1 (White) or 2 (non-white); household income = household income in thousands of dollars when child is 9 years of age; parental knowledge = child’s evaluation of parental knowledge of peer associations at age 9; close to mother = child’s rated closeness to mother at age 9; routine activities = unsupervised routine activities with peers assessed at age 9; peer delinquency-9 = peer delinquency at age 9 as rated by mother; peer delinquency-15 = peer delinquency at age 15 as rated by the child; child delinquency-9 = self-report of child’s own delinquency at age 9; child delinquency-15 = self-report of child’s own delinquency at age 15.

†

p < .0014 (Bonferroni-corrected α level; .05 / 36 correlations).

Main Analyses

The results of a two-regression equation analysis of the male sample revealed significant peer influence (peer delinquency → child delinquency) and peer selection (child delinquency → peer delinquency) effects, a significant facilitative effect for unstructured routine activities, and a nonsignificant inhibitory effect for closeness to mother in both models (see Table 3 and Figure 1). The R² for the equation predicting child delinquency-15 in males was .08 and the R² for the equation predicting peer delinquency-15 in males was .03.

Table 3.

Results of a Linear Regression Analysis of Child Delinquency and Peer Delinquency at Age 15: Male Participants.

Predictor	b [95% CI]	β	z	p
Child delinquency-15 (outcome)
Race	0.458 [0.188, 0.729]	0.067	3.32	<.001
Household Income	−0.003 [–0.005, –0.001]	−0.062	−3.21	.001
Parental knowledge	−0.045 [–0.165, 0.074]	−0.019	−0.74	.458
Close to mother	0.011 [–0.197, 0.219]	0.003	0.10	.918
Routine activities	0.130 [0.043, 0.216]	0.074	2.94	.003
Peer delinquency-9	0.558 [0.158, 0.958]	0.083	2.73	.006
Child delinquency-9	0.269 [0.176, 0.363]	0.189	5.65	<.001
Peer delinquency-15 (outcome)
Race	0.237 [0.070, 0.405]	0.060	2.78	.006
Household income	−0.001 [–0.002, 0.000]	−0.041	−2.22	.026
Parental knowledge	−0.032 [–0.104, 0.039]	−0.023	−0.89	.372
Close to mother	−0.040 [–0.147, 0.066]	−0.020	−0.74	.456
Routine activities	0.065 [0.014, 0.116]	0.064	2.52	.012
Child delinquency-9	0.081 [0.031, 0.131]	0.098	3.18	.001
Peer delinquency-9	0.212 [–0.050, 0.474]	0.054	1.58	.113
Peer-9 with delinquency-9	0.140 [0.097, 0.182]	0.199	6.40	<.001
Peer-15 with delinquency-15	1.894 [1.415, 2.373]	0.513	7.75	<.001

Note. Child delinquency-15 (outcome) = regression equation with child delinquency at age 15 as the outcome measure; peer delinquency-15 (outcome) = regression equation with peer delinquency at age 15 as the outcome measure; with = covariance; race = 1 (White) or 2 (non-White); household income = household income in thousands of dollars when child was 9 years of age; parental knowledge = child’s evaluation of parental knowledge at age 9; close to mother = child’s rated closeness to mother at age 9; routine activities = unsupervised routine activities with peers measured at age 9; peer delinquency-9 = peer delinquency at age 9 as rated by mother; child delinquency-9 = self-report of child’s own delinquency at age 9; b [95% CI] = unstandardized coefficient and the lower and upper limits of the 95% CI for the unstandardized coefficient (in brackets); β = standardized coefficient; z = Wald Z-test statistic; p = significance level of the Wald Z-test statistic; N = 1,759. CI = confidence interval.

Figure 1.

Results of a two-equation maximum likelihood linear regression analysis in which child delinquency, routine activities, closeness to mother, and peer delinquency at age 9 were used to predict child delinquency and peer delinquency at age 15 in male participants.

Regression results in the female sample disclosed a significant peer selection effect, a borderline significant peer influence effect (p = .06), and a significant inhibitory effect for closeness to the mother, coupled with a nonsignificant facilitative effect for unstructured routine activities (see Table 4 and Figure 2). The R² for the equation predicting child delinquency-15 in females was .08 and the R² for the equation predicting peer delinquency-15 was .06.

Table 4.

Results of a Linear Regression Analysis of Child Delinquency and Peer Delinquency at Age 15: Female Participants.

Predictor	b [95% CI]	β	z	p
Child delinquency-15 (outcome)
Race	0.176 [–0.085, 0.437]	0.036	1.32	.186
Household income	−0.004 [–0.006, –0.002]	−0.113	−4.13	<.001
Parental knowledge	0.098 [0.005, 0.191]	0.053	2.06	.039
Close to mother	−0.220 [–0.396, –0.044]	−0.086	−2.44	.015
Routine activities	0.016 [–0.042, 0.075]	0.013	0.55	.582
Peer delinquency-9	0.439 [–0.022, 0.899]	0.071	1.87	.062
Child delinquency-9	0.261 [0.142, 0.380]	0.183	4.31	<.001
Peer delinquency-15 (outcome)
Race	0.096 [–0.081, 0.272]	0.028	1.06	.290
Household income	−0.002 [–0.004, 0.000]	−0.057	−1.49	.136
Parental knowledge	0.088 [0.028, 0.148]	0.068	2.89	.004
Close to mother	−0.119 [–0.229, –0.009]	−0.066	−2.12	.034
Routine activities	0.020 [–0.023, 0.063]	0.023	0.91	.364
Child delinquency-9	0.174 [0.110, 0.238]	0.171	5.35	<.001
Peer delinquency-9	0.309 [–0.065, 0.682]	0.071	1.62	.105
Peer-9 with delinquency-9	0.072 [0.040, 0.105]	0.186	4.33	<.001
Peer-15 with delinquency-15	0.961 [0.664, 1.258]	0.425	6.34	<.001

Figure 2.

To test whether the differences between effects were quantitative, four sets of regression coefficients were compared using a z-test procedure proposed by Paternoster, Brame, Mazerolle, and Piquero (1998). Evaluating the results with a one-tailed z test—given that the relationships were predicted a priori—unstructured routine activities were significantly more predictive of delinquency in boys than in girls, z = 2.15, p < .05, although there was no significant sex difference for peer delinquency, z = 1.32, p = .09. Maternal closeness, however, did a significantly better job of protecting girls than boys against future child, z = −1.66, p < .05, and peer, z = −2.06, p < .05, delinquency.

Sensitivity Testing

Although most of the variables in this study were measured with a single item, both outcome variables were measured with multiple-item scales. A sensitivity or robustness check was performed by creating latent variables for the three multiple-item measures (Wave 5 child delinquency, Wave 6 child delinquency, and Wave 6 peer delinquency) and four single-item measures (Wave 5 maternal closeness, Wave 5 routine activities, Wave 5 parental knowledge, and Wave 5 peer delinquency). Items were loaded onto a latent factor to create latent variables from the three multiple-item measures and a single item was loaded onto a latent factor, after fixing the residual variance to 0 and the factor loading to 1, to create latent variables from the four single-item measures. In the male sample, the use of latent variables resulted in a significant routine activity effect for child (z = 2.29, p < .05) and peer (z = 2.44, p < .05) delinquency, and a nonsignificant maternal closeness effect for child (z = −0.08, p = .94) and peer (z = −0.58, p = .56) delinquency. In the female sample, the use of latent outcome variables led to a significant maternal closeness effect for child (z = −3.05, p < .01) and peer (z = −2.72, p < .01) delinquency, and a nonsignificant routine activity effect for child (z = −0.00, p = .99) and peer (z = 0.62, p = .53) delinquency.

Discussion

The first hypothesis tested in this study held that the peer influence and selection effects would be significant in both the male and female 15-year-olds. Mixed results were obtained. Although the peer influence and peer selection effects were both significant in the male sample, only the peer selection effect was significant in the female sample. Although the absence of a peer influence effect in the female sample was inconsistent with the first hypothesis, it is consistent with several studies showing that peers have a weaker impact on female offending than they do on male offending (Augustyn & McGloin, 2013; Fagan, Van Horn, Hawkins, & Arthur, 2007; Novak & Crawford, 2010; Walters, 2014). The second hypothesis tested in this study stated that maternal closeness would serve a significantly stronger protective effect for girls than boys, and involvement in unsupervised routine activities would place boys at significantly greater risk than girls, and that they would do so independent of the peer influence and peer selection effects. Results from the two regression analyses were fully congruent with this hypothesis, in that, unsupervised routine activities in boys at age 9 predicted higher levels of delinquency at age 15, and maternal closeness in girls at age 9 predicted lower levels of delinquency at age 15. These results mirror those obtained in a study on African American male and female adolescents, where maternal involvement and monitoring protected girls but not boys from future delinquent involvement, and delinquent peers were predictive of future offending in boys but not girls (Bowman, Prelow, & Weaver, 2007). These results are inconsistent, however, with several studies that failed to identify gender differences in maternal support (Silverman & Caldwell, 2005) or peer delinquency (Jennings, Maldonado-Molina, & Komro, 2010; Reitz et al., 2007).

The current results offer preliminary support for the gendered pathways approach to crime, which maintains that male and female pathways to crime are both similar and different (Chesney-Lind & Palko, 2004). As has been reported previously in the literature (Caravita, Sijtsema, Rambaran, & Gini, 2014; Weerman, 2011), the peer selection effect was significant in both the male and female samples of FFCWS participants, although the peer influence effect was only significant in the male sample. Moreover, clear differences were noted in the social environmental factors that accompanied the peer influence and selection effects that formed the male and female pathways to child and peer delinquency. Whereas boys were encouraged to commit crime and associate with delinquent peers through their involvement in unsupervised routine activities, girls were discouraged from engaging in crime and associating with delinquent peers by feelings of closeness to their mothers. The social environmental factor that served as a control variable in this study—perceived parental knowledge of friendship networks—failed to achieve an effect in the male sample and encouraged, rather than discouraged, future child and peer delinquency in the female sample. The lack of effect for perceived parental knowledge in males could perhaps be explained by the results of an earlier study in which perceived parental knowledge failed to predict delinquency directly but did achieve an indirect effect, by reducing male youths’ involvement in unstructured routine activities (Walters, 2018b) or it could be the result of the 6-year time lag between waves. The presence of a risk rather than protective effect for parental knowledge in the female sample could reflect a belief on the part of the girls that their parents were being overbearing and intrusive (Hawk et al., 2013; Kerr & Stattin, 2000).

Implications

Whereas none of the equations in the current investigation predicted more than 8% of the variance in child or peer delinquency, there are still practical implications to these results. One such implication is that the peer selection effect is as strong in females as it is in males and should be assessed in at-risk youth regardless of sex. A second practical implication is that males and females appear to respond differentially to social environmental factors other than peer influence and selection, such that these factors serve important risk and protective functions for future child and peer delinquency. I refer specifically to the effects of unstructured routine activities and maternal closeness on long-term changes in the delinquent behavior of a child and the delinquency of his or her peer group. One form of intervention that may be more effective with adolescent males than with adolescent females is the use of structured and adult-supervised after-school and weekend activities, to reduce the amount of unstructured time boys have to interact with their friends. Even after peer delinquency has been controlled, either in this study or in several previous investigations (Augustyn & McGloin, 2013; McGloin & Shermer, 2009), unsupervised routine activities meaningfully predict delinquent behavior, at least in boys. Strengthening the mother‒daughter relationship is an intervention that may be more effective with adolescent girls than with adolescent boys considering that its protective effect may be limited to girls, as observed in this study and several previous investigations (Bowman et al., 2007; Hill & Atkinson, 1988; Walters, 2013).

Unstructured routine activities in boys and maternal closeness in girls appeared to affect child and peer delinquency equally. In other words, male involvement in unstructured routine activities predicted child delinquency above and beyond the contributions of peer influence, and predicted peer delinquency above and beyond the contributions of peer selection. By the same token, maternal closeness inhibited female involvement in future delinquent behavior net the effects of peer influence, and inhibited female association with delinquent peers net the effects of peer selection. This would seem to suggest that some of the antecedent conditions that promote or inhibit child delinquency also promote or inhibit peer delinquency. There are at least two connotations to these results. First, they imply that child and peer delinquency may share some of the same risk and protective factors. What is required next is research into whether these similarities highlight meaningful commonalities in antecedents to child and peer delinquency or whether they simply reflect the possibility that adolescents project their own level of delinquency onto their friends (Young, Rebellon, Barnes, & Weerman, 2013). Second, shared antecedent conditions for child and peer delinquency may indicate that child and peer delinquency may be equally responsive to certain forms of intervention.

Whereas determining whether child and peer delinquency respond to some of the same interventions is important, so too is ascertaining whether male and female offenders respond to some of the same assessment and intervention protocols. In the previously mentioned Reisig et al. (2006) study, it was noted that the LSI-R was significantly better at identifying criminal risk in adult male offenders than it was at identifying criminal risk in adult female offenders. Examining this same issue in a meta-analysis of studies assessing risk in male and female juvenile offenders, Pusch and Holtfreter (2018) found greater support for gender-neutral risk assessment. This raises a question of whether and how much assessment and treatment should be tailored to participant sex to best meet the individual needs of male and female offenders. Although cognitive–behavioral therapy has been found to reduce recidivism in male and female juvenile offenders (Landenberger & Lipsey, 2005), and peer resistance training, a procedure grounded in cognitive–behavioral and social learning principles, is known to reduce violence and drug use in both boys and girls (Botvin, Griffin, & Nichols, 2006), there is still a great deal we do not know. Additional research is, consequently, required to determine the best way to assess risk in male and female offenders and the optimal treatment approaches for male and female adult and juvenile offenders as a means of reducing both child delinquency and delinquent peer associations.

Limitations

As is true of any study, the current investigation suffered from several limitations. First, the social environmental predictor variables were each measured with a single item. Although the items appeared to do an adequate job of capturing the theoretical concepts they were designed to represent, whenever a single item is used to assess a construct, there is always the possibility that idiosyncratic responding on the item could adversely affect the results. Furthermore, single-item indices make poor latent variables (Cole & Preacher, 2014), even though they can and were used to construct latent variables in the sensitivity testing analyses for this study. Research, nonetheless, indicates that under some circumstances, single items may be as effective as scales for the purposes of prediction (Hoeppner, Kelly, Urbanoski, & Slaymaker, 2011; Smith, Davis, Mendoza, & Zhang, 2017). A further limitation of this study is that the Waves 5 and 6 measures of child and peer delinquency were far from identical. Wave 5 peer delinquency, for instance, was a single item rated by the mother, whereas Wave 6 peer delinquency was based on a five-item scale completed by the child. The child delinquency measures from Waves 5 and 6 were both completed by the child but shared only about half their items in common. Furthermore, each was based on a different rating scale. As a consequence, the “repeated measure” correlations for peer delinquency were nonsignificant and the “repeated measure” correlations for child delinquency were significant but only modest in magnitude.

Another potential criticism of this study is that the large sample sizes could have produced trivial effects, seeing as none of the equations accounted for more than 8% of the variance in an outcome measure. This occurred even though outside of basic demographic measures (race, family income), there was only one control variable (parental knowledge) included in the analyses. Such potentially important criminological constructs as low self-control (Gottfredson & Hirschi, 1990), general strain (Agnew, 1992), and callousness–unemotional traits (Frick & White, 2008), which perhaps should have been controlled, were not, because they were omitted from the FFCWS. It could be argued then that low self-control, which tends to be less prominent in females than in males (LaGrange & Silverman, 1999), may have been partially responsible for the gender-differentiated relationships observed in the current study. General strain has, similarly, been found to affect males and females differentially when used to predict deviant behavior (Posick, Farrell, & Swatt, 2013). Most recently, Meldrum, Piquero, Ozkan, and Powell (2018) determined that remorselessness had a greater effect on youth offending in males than on youth offending in females. In future research, investigators may want to consider controlling for the effects of low self-control, general strain, and callous–unemotional traits, when studying the effects of gender on maternal closeness and unstructured routine activities prediction of child and peer delinquency.

One could argue that the definition or model of peer selection used in this study (child delinquency → peer delinquency) did a poor job of capturing the construct it was designed to assess. Gottfredson and Hirschi’s (1990) definition of selection covers a number of behaviors and activities not included in the peer selection model employed in the current study, so it too could be viewed as a significant study limitation. Furthermore, the 6-year gap between Waves 5 and 6 of the FFCWS is potentially problematic, in that, it leaves open the possibility that unmeasured covariate confounders could have intervened between the independent (maternal closeness, unstructured routine activities) and dependent (peer and child delinquency) variables. I would argue that some of these confounders were actually important mediating variables that served as links between maternal closeness and unsupervised routine activities at age 9 and child and peer delinquency at age 15; this may be why the independent variables were able to predict the dependent variables over such a long span of time. What needs to be kept in mind, however, is that despite the use of single-item measures, nonidentical delinquency indicators, a limited definition of selection, a 6-year gap between waves, large samples, and a modicum of control variables, all eight pathways were subjected to the same less-than-ideal methodological conditions. And, as such, only the predicted pathways achieved significance, and in three out of four cases, the predicted pathway was significantly stronger than the control pathway.

Conclusion

The peer influence and selection effects were observed in male and female adolescents from the FFCWS based on evaluations occurring over a 6-year period. Child and peer delinquency measured in late childhood were found to predict changes in peer and child delinquency assessed from late childhood to midadolescence. Although interesting, these relationships were not the focus of the current investigation. Rather, it was the role of other social environmental factors (i.e., unstructured routine activities and maternal closeness) in predicting a change in child and peer delinquency above and beyond the contributions of peer influence and peer selection that defined the research question and hypotheses for this study. Unlike the peer influence and selection effects, which produced fairly uniform across-sex results in the current study, the effects of the other social environmental variables differed as a function of whether male or female data were being analyzed. There is evidence from previous research that delinquent behavior in boys is more heavily influenced by peers than delinquent behavior in girls (Fagan et al., 2007; Piquero, Gover, MacDonald, & Piquero, 2005; Walters, 2014), and that closeness to one’s mother has a stronger bearing on delinquent behavior in girls than it does on delinquent behavior in boys (Bowman et al., 2007; Hill & Atkinson, 1988; Walters, 2013). The current results support these patterns and suggest that the gendered pathways theory of crime has value as a way of both conceptualizing and altering delinquent behavior and criminal justice policy (Wattanaporn & Holtfreter, 2014). From these results, I would surmise that social environmental factors assessed at age 9 can have an important bearing on gendered differences in delinquent and peer patterns 6 years later.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Glenn D. Walters

Author Biography

Glenn D. Walters, PhD, is a professor in the Department of Criminal Justice at Kutztown University where he teaches classes in criminology, corrections, research methods, and substance misuse and crime. His research interests include peer and parenting effects and delinquency and criminal thinking.

References

Agnew

(1992). Foundation for a general strain theory of crime and delinquency. Criminology, 30, 47-87.

Akers

R. L.

(1998). Social learning and social structure: A general theory of crime and deviance. Boston, MA: Northeastern University Press.

Allison

P. D.

(2012). Handling missing data by maximum likelihood (SAS Global Forum 2012, Paper 312-2012). Cary, NC: NCSAS Institute.

Alvarez-Rivera

L. L.

(2016). Maternal attachment, self-control, and offending among college students. Journal of Ethnicity in Criminal Justice, 14, 58-75.

Andrews

D. A.

Bonta

(1995). The Level of Service Inventory-Revised. Toronto, Ontario, Canada: Multi-Health Systems.

Augustyn

M. B.

McGloin

J. M.

(2013). The risk of informal socializing with peers: Considering gender differences across predatory delinquency and substance use. Justice Quarterly, 30, 117-143.

Belknap

(2007). The invisible woman: Gender, crime, and justice. Belmont, CA: Wadsworth.

Belknap

Holsinger

(2006). The gendered nature of risk factors for delinquency. Feminist Criminology, 1, 48-71.

Bloom

Owen

Rosenbaum

Deschenes

E. P.

(2003). Focusing on girls and young women: A gendered perspective on female delinquency. Women and Criminal Justice, 14, 117-136.

10.

Botvin

G. J.

Griffin

K. W.

Nichols

T. D.

(2006). Preventing youth violence and delinquency through a university school-based prevention approach. Prevention Science, 7, 403-408.

11.

Bowman

M. A.

Prelow

H. M.

Weaver

S. R.

(2007). Parenting behaviors, association with deviant peers, and delinquency in African American adolescents: A mediated-moderation model. Journal of Youth and Adolescence, 36, 517-527.

12.

Brennan

Breitenbach

Dieterich

Salisbury

E. J.

Van Voorhis

(2012). Women’s pathways to serious and habitual crime: A person-centered analysis incorporating gender responsive factors. Criminal Justice and Behavior, 39, 1481-1508.

13.

Broidy

Payne

Piquero

A. R.

(2018). Making sense of heterogeneity in the influence of childhood abuse, mental health, and drug use on women’s offending patterns. Criminal Justice and Behavior, 45, 1565-1587.

14.

Byrne

C. F.

Trew

K. J.

(2008). Pathways through crime: The development of crime and desistance in the accounts of men and women offenders. Howard Journal of Criminal Justice, 47, 238-258.

15.

Caravita

S. C. S.

Sijtsema

J. J.

Rambaran

A. J.

Gini

(2014). Peer influences on moral disengagement in late childhood and early adolescence. Journal of Youth and Adolescence, 43, 193-207.

16.

Chesney-Lind

Palko

(2004). The female offender: Girls, women, and crime (2nd ed.). Thousand Oaks, CA: Sage.

17.

Cole

D. A.

Maxwell

S. E.

(2003). Testing mediational models with longitudinal data: Questions and tips in the use of structural equation modeling. Journal of Abnormal Psychology, 112, 558-577.

18.

Cole

D. A.

Preacher

K. J.

(2014). Manifest variable path analysis: Potentially serious and misleading consequences due to uncorrected measurement error. Psychological Methods, 19, 300-315.

19.

Collins

L. M.

Schafer

J. L.

Karn

C.-M.

(2001). A comparison of inclusive and restrictive strategies in modern missing data procedures. Psychological Methods, 6, 330-351.

20.

Cutrin

Gómez-Fraguela

J. A.

Sobral

(2017). Two faces of parental support: Risk and protection for antisocial youth depending on parental knowledge. Journal of Child and Family Studies, 26, 296-305.

21.

Daly

(1992). Women’s pathways to felony court: Feminist theories of lawbreaking and problems of representation. Southern California Review of Law and Women’s Studies, 2, 11-52.

22.

Fagan

A. A.

Van Horn

M. L.

Hawkins

J. D.

Arthur

M. W.

(2007). Gender similarities and differences in the association between risk and protective factors and self-reported serious delinquency. Prevention Science, 8, 115-124.

23.

Frick

P. J.

White

S. F.

(2008). Research review: The importance of callous-unemotional traits for developmental models of aggression and antisocial behavior. Child Psychology and Psychiatry, 49, 359-375.

24.

Gottfredson

M. R.

Hirschi

(1990). A general theory of crime. Stanford, CA: Stanford University Press.

25.

Hawk

S. T.

Keijsers

Frijns

Hale

W. W.

Branje

Meeus

(2013). “I still haven’t found what I’m looking for”: Parental privacy invasion predicts reduced parental knowledge. Developmental Psychology, 49, 1286-1298.

26.

Higgins

G. E.

Jennings

W. G.

Mahoney

(2010). Developmental trajectories of maternal and paternal attachment and delinquency in adolescence. Deviant Behavior, 31, 655-677.

27.

Hill

G. D.

Atkinson

M. P.

(1988). Gender, familial control, and delinquency. Criminology, 26, 127-149.

28.

Hoeben

E. M.

Weerman

F. M.

(2016). Why is involvement in unstructured socializing related to adolescent delinquency? Criminology, 54, 242-281.

29.

Hoeppner

B. B.

Kelly

J. F.

Urbanoski

K. A.

Slaymaker

(2011). Comparative utility of a single-item vs. multiple-item measure of self-efficacy in predicting relapse among young adults. Journal of Substance Abuse Treatment, 41, 305-312.

30.

Hughes

L. A.

Short

J. F.

(2014). Partying, cruising, and hanging in the streets: Gangs, routine activities, and delinquency and violence in Chicago, 1959-1962. Journal of Quantitative Criminology, 30, 415-451.

31.

Jennings

W. G.

Maldonado-Molina

M. M.

Komro

K. A.

(2010). Sex similarities/differences in trajectories of delinquency among urban Chicago youth: The role of delinquent peers. American Journal of Criminal Justice, 35, 56-75.

32.

Jones

N. J.

Brown

S. L.

Wanamaker

K. A.

Greiner

L. E.

(2014). A quantitative exploration of gendered pathways to crime in a sample of male and female juvenile offenders. Feminist Criminology, 9, 113-136.

33.

Kerr

Stattin

(2000). What parents know, how they know it, and several forms of adolescent adjustment: Further support for a reinterpretation of monitoring. Developmental Psychology, 36, 366-380.

34.

Kiesner

Poulin

Nicotra

(2003). Peer relations across contexts: Individual-network homophily and network inclusion in and after school. Child Development, 74, 1328-1343.

35.

Kruttschnitt

(2013). Gender and crime. Annual Review of Sociology, 39, 291-308.

36.

LaGrange

T. C.

Silverman

R. A.

(1999). Low self-control and opportunity: Testing the general theory of crime as an explanation for gender differences in delinquency. Criminology, 37, 41-72.

37.

Landenberger

N. A.

Lipsey

M. W.

(2005). The positive effects of cognitive-behavioral programs for offenders: A meta-analysis of factors associated with effective treatment. Journal of Experimental Criminology, 1, 451-476.

38.

McGloin

J. M.

Shermer

L. O.

(2009). Self-control and deviant peer network structure. Journal of Research in Crime & Delinquency, 46, 35-72.

39.

Meldrum

R. C.

Piquero

A. R.

Ozkan

Powell

Z. A.

(2018). An examination of the criminological consequences and correlates of remorselessness during adolescence. Youth Violence and Juvenile Justice, 16, 279-298.

40.

Moffitt

T. E.

(1993). Adolescence-limited and life-course-persistent antisocial behavior: A developmental taxonomy. Psychological Review, 100, 674-701.

41.

Monahan

K. C.

Steinberg

Cauffman

(2009). Affiliation with antisocial peers, susceptibility to peer influence, and antisocial behavior during the transition to adulthood. Developmental Psychology, 45, 1520-1530.

42.

Muthén

. (1997-2017). Mplus user’s guide (8th ed.). Los Angeles, CA: Author.

43.

Novak

K. B.

Crawford

L. A.

(2010). Routine activities as determinants of gender differences in delinquency. Journal of Criminal Justice, 38, 913-920.

44.

Osgood

D. W.

Wilson

J. K.

O’Malley

P. M.

Bachman

J. G.

Johnston

L. D.

(1996). Routine activities and individual deviant behavior. American Sociological Review, 61, 635-655.

45.

Paternoster

Brame

Mazerolle

Piquero

(1998). Using the correct statistical test for the equality of regression coefficients. Criminology, 36, 859-866.

46.

Peyre

Leplége

Coste

(2011). Missing data methods for dealing with missing items in quality of life questionnaires: A comparison by simulation of personal mean score, full information maximum likelihood, multiple imputation, and hot deck techniques applied to the SF-36 in the French 2003 decennial health survey. Quality of Life Research, 20, 287-300.

47.

Piquero

N. L.

Gover

A. R.

MacDonald

J. M.

Piquero

A. R.

(2005). The influence of delinquent peers on delinquency: Does gender matter? Youth and Society, 36, 251-275.

48.

Posick

Farrell

Swatt

M. L.

(2013). Do boys fight and girls cut? A general theory approach to gender and deviance. Deviant Behavior, 34, 685-705.

49.

Pusch

Holtfreter

(2018). Gender and risk assessment in juvenile offenders: A meta-analysis. Criminal Justice and Behavior, 45, 56-81.

50.

Reichman

N. E.

Teitler

J. O.

Garfinkel

McLanahan

(2001). Fragile families: Sample and design. Children & Youth Services Review, 23, 303-326.

51.

Reisig

M. D.

Holtfreter

Morash

(2006). Assessing recidivism risk across female pathways to crime. Justice Quarterly, 23, 384-405.

52.

Reitz

Prinzie

Dekovíc

Buist

(2007). The role of peer contacts in the relationship between parental knowledge and adolescents’ externalizing behaviors: A latent growth curve modeling approach. Journal of Youth and Adolescence, 36, 623-634.

53.

Salisbury

E. J.

Van Voorhis

(2009). Gendered pathways: A quantitative investigation of women probationers’ paths to incarceration. Criminal Justice and Behavior, 36, 541-566.

54.

Schroeder

R. D.

Higgins

G. E.

Mowen

T. J.

(2014). Maternal attachment trajectories and criminal offending by race. American Journal of Criminal Justice, 39, 155-171.

55.

Sedding

(2014). Peer group association, the acceptance of norms and violent behaviour: A longitudinal analysis of reciprocal effects. European Journal of Criminology, 11, 319-339.

56.

Silverman

J. R.

Caldwell

R. M.

(2005). The influence of parental emotional support and monitoring on self-reported delinquent impulsive behavior and noncompliance among juvenile offenders: An examination of gender differences. International Journal of Forensic Mental Health, 4, 159-174.

57.

Simpson

S. S.

Yahner

J. L.

Dugan

(2008). Understanding women’s pathways to jail: Analysing the lives of incarcerated women. Australian & New Zealand Journal of Criminology, 41, 84-108.

58.

Smith

D. C.

Davis

J. P.

Mendoza

Zhang

(2017). Predictive validity of single-item motivational measures for emerging adults in substance-use disorder treatment. Journal of the Society for Social Work and Research, 8, 259-277.

59.

Svensson

Burk

W. J.

Stattin

Kerr

(2012). Peer selection and influence of delinquent behavior of immigrant and nonimmigrant youths: Does context matter? International Journal of Behavioral Development, 36, 178-185.

60.

Turanovic

J. J.

Reisig

M. D.

Pratt

T. C.

(2015). Risky lifestyles, low self-control, and violent victimization across gendered pathways to crime. Journal of Quantitative Criminology, 31, 183-206.

61.

Vieno

Nation

Pastore

Santinello

(2009). Parenting and antisocial behavior: A model of the relationship between adolescent self-disclosure, parental closeness, parental control, and adolescent antisocial behavior. Developmental Psychology, 45, 1509-1519.

62.

Walters

G. D.

(2013). Delinquency, parental involvement, early adult criminality, and sex: Evidence of moderated mediation. Journal of Adolescence, 36, 777-785.

63.

Walters

G. D.

(2014). Sex as a moderator and perceived peer pressure as a mediator of the externalizing-delinquency relationship: A test of gendered pathways theory. Journal of Criminal Justice, 42, 299-305.

64.

Walters

G. D.

(2018a). Mothers and fathers, sons and daughters: Parental knowledge and quality of the parent-child relationship as predictors delinquency in same- and cross-sex parent-child dyads. Manuscript submitted for publication.

65.

Walters

G. D.

(2018b). Unsupervised routine activities as a mediator of the parental knowledge-delinquency relationship. Justice Quarterly, 35, 871-894.

66.

Wang

Kim

S. Y.

Anderson

E. R.

Chen

C.-C.

Yan

(2012). Parent-child acculturation discrepancy, perceived parental knowledge, peer deviance, and adolescent delinquency in Chinese immigrant families. Journal of Youth and Adolescence, 41, 907-919.

67.

Warr

(2002). Companions in crime: The social aspects of criminal conduct. Cambridge, UK: Cambridge University Press.

68.

Wattanaporn

K. A.

Holtfreter

(2014). The impact of feminist pathways research on gender-responsive policy and practice. Feminist Criminology, 9, 191-207.

69.

Weerman

F. M.

(2011). Delinquent peers in context: A longitudinal network analysis of selection and influence effects. Criminology, 49, 253-286.

70.

Young

J. T. N.

Rebellon

C. J.

Barnes

J. C.

Weerman

F. M.

(2013). Unpacking the black box of peer similarity in deviance: Understanding the mechanisms linking personal behavior, peer behavior, and perceptions. Criminology, 52, 60-86.